Teaching optimization scoring method and system based on large model
Through the large model, the feature extraction and quantitative evaluation of teaching design cases is solved, the problem of traditional teaching design relying on experience is achieved, scientific and personalized teaching optimization is achieved, and teaching quality and student learning effect are improved.
Patent Information
- Application Number
- CN202510571908.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-06
- Publication Date
- 2025-08-15
AI Technical Summary
Traditional teaching design relies too much on teachers' personal experience and intuition, lacks consistency judgment standards, and young teachers and teacher students lack in-depth and systematic thinking on the textbooks, resulting in a lack of scientificity and personalization of teaching design.
By collecting teaching design cases, using big models for feature extraction and multiple rounds of manual compliance, formulating quantifiable evaluation standards, training teaching design scoring models, and providing personalized teaching guidance and resource generation.
A more scientific, accurate and personalized teaching design has been achieved, the teaching quality and student learning efficiency have been improved, the burden on teachers has been reduced, and the teaching interactivity and fun has been enhanced.
Smart Images
Figure CN120494615A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of large models, and specifically provides a teaching optimization scoring method and system based on large models. Background Art
[0002] In this technology, large-scale model technology is combined with the field of education, and has been widely used and achieved remarkable results in the fields of teaching and research.
[0003] With the rapid development of artificial intelligence (AI), big models, as a core driving force in the field, have achieved breakthroughs in areas such as natural language processing and computer vision. In education, big model technology, by deeply mining student learning data and logs, can accurately analyze students' knowledge mastery, learning habits, and interests. This provides comprehensive support for teaching resource development, teaching process optimization, student situation analysis, and teaching management decision-making.
[0004] Traditional instructional design often relies on teachers' personal experience and intuition, lacking scientificity and precision. However, large-scale model-based instructional design optimization and scoring technology leverages the data analysis capabilities and machine learning algorithms of large models to quantitatively evaluate and optimize instructional design. This technology collects and analyzes large amounts of instructional data, including student learning behaviors, learning outcomes, and instructional duration, to construct accurate student profiles and learning path planning. Based on this, the technology can dynamically adjust instructional content and strategies based on student learning progress and feedback, ensuring that every knowledge point is effectively imparted and improving overall instructional quality.
[0005] Furthermore, large-scale model-based instructional design optimization and scoring technology can also enable personalized instruction. By analyzing students' learning data, this technology can identify their individual learning needs and interests, providing them with tailored learning resources and recommendations. This personalized teaching approach not only stimulates students' interest and motivation, but also improves their learning efficiency and academic performance.
[0006] However, existing technologies have the problem that traditional teaching design relies too much on teachers' personal experience and intuition, lacks relatively consistent evaluation criteria, and young teachers and normal school students lack in-depth understanding of teaching materials and lack systematic thinking and systematic consideration. Summary of the Invention
[0007] The present invention aims to address the deficiencies of the above-mentioned prior art and provides a practical teaching optimization scoring method based on a large model.
[0008] A further technical task of the present invention is to provide a teaching optimization scoring system based on a large model that is rationally designed, safe and applicable.
[0009] The technical solution adopted by the present invention to solve its technical problem is:
[0010] A teaching optimization scoring method based on a large model has the following steps:
[0011] S1. Collect a certain number of teaching design cases and pre-process the design cases;
[0012] S2. Use the large model to extract features from the pre-processed teaching design cases to obtain the feature vector of each case. Through multiple rounds of manual matching of the feature vectors, the key evaluation points in the teaching design are clarified.
[0013] S3. Label scores for a large number of teaching design data sets, train teaching design scoring models, and predict scores based on the feature vectors of teaching designs to provide guidance and assistance to young teachers and normal school students.
[0014] Furthermore, in step S1, based on the teaching design optimization scoring technology of the big model, the data analysis capabilities and machine learning algorithms of the big model are used to quantitatively evaluate and optimize the collected teaching designs;
[0015] By collecting and analyzing a large amount of teaching data, we construct accurate student portraits and learning path planning, and further dynamically adjust the collected teaching content and teaching strategies based on students' learning progress and feedback, so that they are effectively marked in the training data set.
[0016] Furthermore, big model technology integrates and analyzes educational resources and student data. In terms of generating educational resources, AI big models are used to generate teaching content based on teachers' teaching needs and students' specific learning situations.
[0017] Furthermore, in step S2, the large amount of teaching design content in education intelligence is broken down into seven parts: curriculum analysis, teaching objectives, evaluation design, key and difficult points of teaching, teaching methods, teaching process, and homework design.
[0018] Based on these seven parts, quantifiable evaluation criteria are formulated. In the process of collecting and annotating teaching design data sets, corresponding evaluation suggestions are generated for these seven parts. The big model evaluates and scores the teaching design according to the disassembled standard scoring points.
[0019] A teaching optimization scoring system based on a large model is characterized by: first, collecting a certain number of teaching design cases and preprocessing the design cases; then, using the large model to extract features of the preprocessed teaching design cases to obtain a feature vector for each case; and through multiple rounds of manual matching of the feature vectors, identifying the key parity points in the teaching design.
[0020] Finally, we annotate the scores of a large number of teaching design data sets, train the teaching design scoring model, and predict the scores of the teaching designs based on their feature vectors, providing guidance and help to young teachers and normal school students.
[0021] Furthermore, after pre-processing, the teaching design optimization scoring technology based on the big model uses the data analysis capabilities and machine learning algorithms of the big model to quantitatively evaluate and optimize the collected teaching designs;
[0022] By collecting and analyzing a large amount of teaching data, we construct accurate student portraits and learning path planning, and further dynamically adjust the collected teaching content and teaching strategies based on students' learning progress and feedback, so that they are effectively marked in the training data set.
[0023] Furthermore, big model technology integrates and analyzes educational resources and student data. In terms of generating educational resources, AI big models are used to generate teaching content based on teachers' teaching needs and students' specific learning situations.
[0024] Furthermore, the key evaluation points in teaching design are clarified, and the large amount of teaching design content in education intelligence is broken down into seven parts: curriculum analysis, teaching objectives, evaluation design, key and difficult points of teaching, teaching methods, teaching process, and homework design.
[0025] Based on these seven parts, quantifiable evaluation criteria are formulated. In the process of collecting and annotating teaching design data sets, corresponding evaluation suggestions are generated for these seven parts. The big model evaluates and scores the teaching design according to the disassembled standard scoring points.
[0026] Compared with the prior art, the teaching optimization scoring method and system based on a large model of the present invention has the following outstanding beneficial effects:
[0027] The present invention can solve the problems that traditional teaching design relies too much on teachers' personal experience and intuition, lacks relatively consistent evaluation criteria, young teachers and normal school students lack in-depth understanding of teaching materials, are restricted by linear thinking and analytical thinking, and lack systematic thinking and systematic consideration. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0029] Attachment Figure 1It is a framework diagram of a teaching optimization scoring method based on a large model. DETAILED DESCRIPTION
[0030] In order to enable those skilled in the art to better understand the solutions of the present invention, the present invention will be further described in detail below in conjunction with specific embodiments. Obviously, the embodiments described are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0031] A best embodiment is given below:
[0032] like Figure 1 As shown, a teaching optimization scoring method based on a large model in this embodiment has the following steps:
[0033] S1. Collect a certain number of teaching design cases and pre-process the design cases;
[0034] This large-scale model-based instructional design optimization and scoring technology leverages the model's data analysis capabilities and machine learning algorithms to quantitatively evaluate and optimize collected instructional designs. By collecting and analyzing large amounts of instructional data, including student learning behaviors, learning outcomes, and instructional duration, it constructs accurate student profiles and learning path planning. Based on this foundation, it dynamically adjusts collected instructional content and strategies based on student learning progress and feedback, ensuring that every knowledge point is effectively labeled in the training dataset and improving overall data quality.
[0035] By integrating and analyzing vast amounts of educational resources and student data through big model technology, teachers can be provided with more scientific and comprehensive teaching support. Regarding the generation of educational resources, AI big models can generate high-quality and diverse teaching content, such as lesson plans, courseware, and exercises, based on teachers' teaching needs and students' specific learning situations. These resources not only enrich teaching methods but also ensure the professionalism, accuracy, and relevance of teaching content.
[0036] Furthermore, large model technology can serve as an intelligent teaching assistant for teachers, providing real-time Q&A and generating teaching resources in the classroom. This not only reduces the teacher's workload but also makes classroom teaching more interactive and engaging. Large model technology also plays a crucial role in after-class practice and independent learning. Through intelligent Q&A and assessment mechanisms, it helps students resolve learning difficulties promptly and improve learning outcomes.
[0037] To sum up, teaching design based on big model technology can break the limitations of traditional teaching that relies too much on teachers' personal experience and intuition, and achieve more scientific, accurate and personalized teaching.
[0038] S2. Use the large model to extract features from the pre-processed teaching design cases to obtain the feature vector of each case. Through multiple rounds of manual matching of the feature vectors, the key evaluation points in the teaching design are clarified.
[0039] By breaking down the vast amount of instructional design content within smart education, traditional instructional design is broken down into seven core components: curriculum analysis, teaching objectives, assessment design, key points and difficulties, teaching methods, teaching process, and assignment design. Quantifiable evaluation criteria are then developed based on these seven core components. Scoring criteria encompass dimensions such as completeness, logic, operability, and innovation to ensure the quality of instructional design. During the collection and annotation of instructional design datasets, corresponding evaluation suggestions are generated for these seven core components, allowing the big model to learn how to guide and evaluate instructional designs. The evaluation and scoring of instructional designs are then performed based on the decomposed standard scoring criteria, enabling the big model to assist teachers in their daily instructional design work and to guide and cultivate students in their learning process. Case studies are then manually scored based on the instructional design scoring criteria, forming a labeled dataset.
[0040] The detailed scoring table for the large-scale model teaching design is as follows:
[0041] To develop a scientific and rational grading rubric, it is necessary to formulate quantifiable evaluation criteria based on seven core components: curriculum analysis, teaching objectives, assessment design, key and difficult points, teaching methods, teaching process, and assignment design. These rubrics should encompass dimensions such as completeness, logic, operability, and innovation to ensure the quality of the instructional design.
[0042] Ideas for setting scoring criteria:
[0043] Use a scale scoring method: such as a 5-point system.
[0044] Refine the scoring items for each section: Ensure that each teaching design link has clear evaluation indicators and can clearly distinguish different levels of performance.
[0045] Focus on the logic and operability of teaching: not only evaluate the completeness of the content, but also examine whether it conforms to teaching rules and is easy to implement.
[0046] 1. When the scoring item is curriculum standard analysis, the total score is 15, which is divided into curriculum standard matching, target coverage, and adaptability. The split score for curriculum standard matching is 0 - not reflected, 1 - partially met, 3 - basically met, and 5 - completely met. The quantifiable judgment criteria are whether the curriculum standard content is clearly extracted and whether it matches the teaching content.
[0047] The target coverage is divided into scores of 0-not reflected, 1-partial coverage, 3-basic coverage, and 5-complete coverage. The quantitative judgment basis is whether the knowledge points and ability requirements required by the curriculum standards are fully covered?
[0048] The adaptability score is divided into 0-no adjustment, 1-unreasonable adjustment, 3-partial adaptation, and 5-high adaptation. The quantifiable judgment basis is whether the teaching content is adjusted in accordance with the students' learning situation?
[0049] 2. When the scoring item is a teaching objective, the total score is 10, and the scoring items are clear objectives and complete objectives;
[0050] The score for goal clarity is 0-unclear, 1-vague, 3-basic clear, 5-completely clear. The quantifiable judgment basis is whether the goal meets the SMART principle (specific, measurable, and achievable)?
[0051] The score for goal completeness is 0 - does not reflect the three-dimensional goals, 1 - partially reflected, 3 - basically complete, 5 - completely complete. The quantifiable judgment basis is whether it covers the three aspects of "knowledge and skills, processes and methods, and emotional attitudes"?
[0052] 3. When the scoring item is evaluation design, the total score is 10. The split scoring items are evaluation diversity and goal matching. The split score for evaluation diversity is 0 - no evaluation design, 1 - single method, 3 - multiple methods, 5 - covering formative and summative methods. The quantifiable judgment basis is whether it includes multiple evaluation methods such as classroom questions, exercises, and tests?
[0053] The target matching score is 0-no match, 1-partial match, 3-basic match, 5-complete match. The quantifiable judgment basis is whether the key knowledge of this lesson is clearly listed and meets the curriculum standards?
[0054] 4. The scoring items are teaching key points and difficulties, with a total score of 15. The sub-scoring items are key point extraction, difficulty analysis and solutions. The sub-scoring items for key point extraction are 0-not extracted, 1-inaccurate extraction, 3-basic accuracy, and 5-highly accurate. The quantifiable judgment basis is whether the key knowledge of this lesson is clearly listed and meets the curriculum standards?
[0055] The difficulty analysis is scored as 0-no analysis, 1-irrational analysis, 3-basically reasonable, and 5-highly reasonable. The quantifiable judgment basis is whether the students' easy mistakes and difficulties in understanding are correctly judged?
[0056] The solution is scored as 0-no solution, 1-untargeted solution, 3-partially effective, and 5-scientifically effective. The quantifiable judgment basis is whether it provides methods to overcome difficulties, such as cases, experiments, and scenarios?
[0057] 5. The scoring item is teaching method, with a total score of 10. The sub-scoring items are rationality of method and interactivity. The sub-scoring items for rationality of method are 0-no method, 1-inappropriate method, 3-basically reasonable, and 5-completely reasonable. The quantifiable judgment basis is whether the teaching method selected is in line with the teaching content?
[0058] The interactive score is 0-no interaction, 1-very little interaction, 3-partial interaction, 5-high interaction. The quantitative judgment is based on whether interactive methods such as group discussion, exploration, and questioning are designed?
[0059] 6. The scoring item is the teaching process, with a total score of 25. The scoring items are complete structure, clear logic, student interaction, practical activities and summary and improvement. The score for complete structure is 0 - missing major links, 1 - partially complete, 3 - basically complete, 5 - complete. The quantifiable judgment basis is whether it includes introduction, new teaching, consolidation, expansion and summary?
[0060] The score for logical clarity is 0-no logic, 1-disordered order, 3-basic clarity, 5-high clarity. The quantifiable judgment basis is whether the classroom process is gradual and naturally transitioned?
[0061] The scores for student interaction are 0-no interactivity, 1-single interaction, 3-diverse interaction, and 5-rich interaction. The quantifiable judgment basis is whether effective classroom questions and answers, discussions, etc. are designed?
[0062] The scores for practical activities are 0-no practical activities, 1-single activity, 3-diverse activities, and 5-rich activities. The quantifiable judgment basis is whether it contains creative and meaningful classroom practical activities?
[0063] The score for summary improvement is 0-no summary, 1-simple summary, 3-basic complete, 5-clear and in-depth. The quantifiable judgment basis is whether the key points are reviewed at the end and extended thinking is proposed?
[0064] 7. The scoring item is homework design, with a total score of 15. The split scoring items are target matching, diverse question types and moderate difficulty. The split score for target matching is 0-no match, 1-partial match, 3-basic match, and 5-complete match. The quantifiable judgment basis is whether the homework is targeted at the core content of the class?
[0065] The scores for various question types are 0-no match, 1-partial match, 3-basic match, 5-complete match. The quantifiable judgment basis is whether it includes basic questions, extension questions, application questions, etc.?
[0066] The score for moderate difficulty is 0-too difficult / too easy, 1-somewhat difficult / somewhat easy, 3-basically moderate, and 5-level adaptation. The quantifiable judgment basis is whether there are different levels of homework to cater to different students?
[0067] S3. Label scores for a large number of teaching design data sets, train teaching design scoring models, and predict scores based on the feature vectors of teaching designs to provide guidance and assistance to young teachers and normal school students.
[0068] At the same time, we can solve the following problems through large models:
[0069] 1. Personalized Learning Path and In-Depth Analysis
[0070] The large model can deeply explore and analyze textbook content, combining the learning styles and interests of young teachers and normal school students to generate personalized learning plans and resources for them. Through detailed textbook analysis, knowledge point expansion, and relevant case studies and examples, it helps them deeply understand the textbook's writing intentions, knowledge structure, and key and difficult teaching points.
[0071] 2. Intelligent Tutoring and Q&A
[0072] The big model can act as an intelligent tutor, monitoring the learning status of young teachers and normal school students in real time and providing timely guidance and answers. Whether it's theoretical questions in the textbook or confusion in teaching practice, the big model can provide accurate answers and detailed explanations, helping them overcome learning obstacles and improve their understanding and application of the textbook.
[0073] 3. Rich and diverse educational resources
[0074] The large model can automatically generate rich and diverse educational content based on teaching needs. This content includes explanations of knowledge points in various subjects, teaching cases, and extension exercises, covering all aspects of the textbook. By incorporating multimedia elements such as images, videos, and audio, the learning process is made more vivid and engaging, helping young teachers and normal school students to more intuitively understand the textbook content.
[0075] 4. Dynamically adjust and optimize teaching plans
[0076] The large model dynamically adjusts and optimizes teaching plans based on the learning progress and feedback of young teachers and teacher trainees. By monitoring their learning outcomes in real time, the large model accurately assesses their understanding of the textbooks and adjusts the difficulty and depth of the teaching content accordingly, ensuring that they consistently improve at a level that suits them.
[0077] Traditional teaching design is often constrained by linear and analytical thinking and lacks systematic thinking and consideration. We use a well-trained teaching design model, which includes professional scoring capabilities for teaching design and a knowledge system of massive basic parameters of the model, to score new teaching design cases, provide scoring details and reasons, and guide users to optimize their teaching designs.
[0078] At the same time, traditional instructional design is often constrained by linear and analytical thinking, lacking systematic thinking and consideration. This leads to the neglect of the complexity of the teaching and learning process, and the lack of integrity and coherence in instructional design. Furthermore, there is a lack of good, easy-to-use instructional design tools, making instructional design the domain of experts and connoisseurs, while frontline teachers find it difficult to conduct effective instructional design. To address this situation, based on the multimodal data processing and fusion analysis capabilities of large models, we provide comprehensive information support for instructional design by collecting student performance, course resources, teacher feedback, and more. Through multi-dimensional analysis, instructional designers can break through the constraints of linear thinking, see the complex connections between different factors, and thus develop more systematic and comprehensive instructional design solutions.
[0079] At the same time, the large model enables real-time assessment of student learning performance, rapid and accurate analysis of student assignments and tests, and targeted feedback and suggestions. This evaluation mechanism not only improves efficiency and accuracy but also helps students understand their learning progress and adjust their learning strategies, embodying the principles of continuous improvement and optimization within systems thinking.
[0080] Finally, the big model has the ability to continuously learn and update, and can constantly adapt to changes in the educational environment and student needs. This ability enables teaching design to keep pace with the times, remain systematic and forward-looking.
[0081] Based on the above method, a teaching optimization scoring system based on a big model in this embodiment is characterized in that, first, a certain number of teaching design cases are collected and the design cases are preprocessed; then, the big model is used to extract features of the preprocessed teaching design cases to obtain a feature vector for each case, and through multiple rounds of manual matching of the feature vectors, the key parity points in the teaching design are clarified.
[0082] Finally, we annotate the scores of a large number of teaching design data sets, train the teaching design scoring model, and predict the scores of the teaching designs based on their feature vectors, providing guidance and help to young teachers and normal school students.
[0083] Among them, after pre-processing, the teaching design optimization scoring technology based on the big model uses the data analysis capabilities and machine learning algorithms of the big model to quantitatively evaluate and optimize the collected teaching designs;
[0084] By collecting and analyzing a large amount of teaching data, we construct accurate student portraits and learning path planning, and further dynamically adjust the collected teaching content and teaching strategies based on students' learning progress and feedback, so that they are effectively marked in the training data set.
[0085] Big model technology integrates and analyzes educational resources and student data. In terms of generating educational resources, AI big models are used to generate teaching content based on teachers' teaching needs and students' specific learning situations.
[0086] Clarify the key evaluation points in teaching design, break down the large amount of teaching design content in smart education, and split the traditional teaching design content into seven parts: curriculum analysis, teaching objectives, evaluation design, key and difficult points of teaching, teaching methods, teaching process, and homework design;
[0087] Based on these seven parts, quantifiable evaluation criteria are formulated. In the process of collecting and annotating teaching design data sets, corresponding evaluation suggestions are generated for these seven parts. The big model evaluates and scores the teaching design according to the disassembled standard scoring points.
[0088] The above-mentioned specific implementation methods are only specific cases of the present invention. The patent protection scope of the present invention includes but is not limited to the above-mentioned specific implementation methods. Any technical solutions that conform to the above-mentioned specific implementation methods of the present invention and any appropriate changes or substitutions made thereto by ordinary technicians in the relevant technical field shall fall within the patent protection scope of the present invention.
[0089] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. A teaching optimization scoring method based on a large model, characterized in that: The steps are as follows: S1. Collect a certain number of teaching design cases and pre-process the design cases; S2. Use the large model to extract features from the pre-processed teaching design cases to obtain the feature vector of each case. Through multiple rounds of manual matching of the feature vectors, the key evaluation points in the teaching design are clarified. S3. Label scores for a large number of teaching design data sets, train teaching design scoring models, and predict scores based on the feature vectors of teaching designs to provide guidance and assistance to young teachers and normal school students.
2. A teaching optimization scoring method based on a large model according to claim 1, characterized in that: In step S1, the teaching design optimization scoring technology based on the big model uses the data analysis capabilities and machine learning algorithms of the big model to quantitatively evaluate and optimize the collected teaching designs; By collecting and analyzing a large amount of teaching data, we construct accurate student portraits and learning path planning, and further dynamically adjust the collected teaching content and teaching strategies based on students' learning progress and feedback, so that they are effectively marked in the training data set.
3. A teaching optimization scoring method based on a large model according to claim 2, characterized in that: Big model technology integrates and analyzes educational resources and student data. In terms of generating educational resources, AI big models are used to generate teaching content based on teachers' teaching needs and students' specific learning situations.
4. A teaching optimization scoring method based on a large model according to claim 3, characterized in that: In step S2, the large amount of teaching design content in education intelligence is broken down into seven parts: curriculum analysis, teaching objectives, evaluation design, key and difficult points of teaching, teaching methods, teaching process, and homework design. Based on these seven parts, quantifiable evaluation criteria are formulated. In the process of collecting and annotating teaching design data sets, corresponding evaluation suggestions are generated for these seven parts. The big model evaluates and scores the teaching design according to the disassembled standard scoring points.
5. A teaching optimization scoring system based on a large model, characterized by: First, a certain number of teaching design cases are collected and preprocessed. Then, the large model is used to extract features from the preprocessed teaching design cases to obtain the feature vectors of each case. Through multiple rounds of manual matching of the feature vectors, the key parity points in the teaching design are identified. Finally, we annotate the scores of a large number of teaching design data sets, train the teaching design scoring model, and predict the scores of the teaching designs based on their feature vectors, providing guidance and help to young teachers and normal school students.
6. A teaching optimization scoring system based on a large model according to claim 5, characterized in that: After pre-processing, the teaching design optimization scoring technology based on the big model uses the data analysis capabilities and machine learning algorithms of the big model to quantitatively evaluate and optimize the collected teaching designs; By collecting and analyzing a large amount of teaching data, we construct accurate student portraits and learning path planning, and further dynamically adjust the collected teaching content and teaching strategies based on students' learning progress and feedback, so that they are effectively marked in the training data set.
7. A teaching optimization scoring system based on a large model according to claim 6, characterized in that: Big model technology integrates and analyzes educational resources and student data. In terms of generating educational resources, AI big models are used to generate teaching content based on teachers' teaching needs and students' specific learning situations.
8. A teaching optimization scoring system based on a large model according to claim 7, characterized in that: Clarify the key evaluation points in teaching design, break down the large amount of teaching design content in smart education, and split the traditional teaching design content into seven parts: curriculum analysis, teaching objectives, evaluation design, key and difficult points of teaching, teaching methods, teaching process, and homework design; Based on these seven parts, quantifiable evaluation criteria are formulated. In the process of collecting and annotating teaching design data sets, corresponding evaluation suggestions are generated for these seven parts. The big model evaluates and scores the teaching design according to the disassembled standard scoring points.